Genetic regulatory network-based symbiotic evolution

► The main idea is not to defeat SE, GA or other algorithms but to introduce a new scheme into evolutionary computation, the gene regulatory network. ► Contrasting first study with third one, by adding GRN with automatically weighted genes in the gene pool, the AR is increased about 82% and the GR is increased about 9%. ► SE and GRNSE are compared for different individual population sizes (M, 2M, and 4M). GRNSE performed better for smaller individual population sizes, which is usually required for hardware constraint and high-speed evolution. ► By studying the inference of various population rates, a range [0.2, 0.6] is recommended for an unknown optimization problem. Most of the functions present a reliable acceleration improvement and an almost better regulatory behavior in this interval.